WO2018001847A1 - Système et procédé de détection automatiques d'images clés - Google Patents

Système et procédé de détection automatiques d'images clés Download PDF

Info

Publication number
WO2018001847A1
WO2018001847A1 PCT/EP2017/065347 EP2017065347W WO2018001847A1 WO 2018001847 A1 WO2018001847 A1 WO 2018001847A1 EP 2017065347 W EP2017065347 W EP 2017065347W WO 2018001847 A1 WO2018001847 A1 WO 2018001847A1
Authority
WO
WIPO (PCT)
Prior art keywords
radiology
image
images
stack
finding
Prior art date
Application number
PCT/EP2017/065347
Other languages
English (en)
Inventor
Amir Mohammad TAHMASEBI MARAGHOOSH
Thusitha Dananjaya De Silva MABOTUWANA
Yuechen Qian
Original Assignee
Koninklijke Philips N.V.
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Koninklijke Philips N.V. filed Critical Koninklijke Philips N.V.
Priority to CN201780046678.5A priority Critical patent/CN109564773B/zh
Priority to US16/314,104 priority patent/US11361530B2/en
Priority to EP17735024.6A priority patent/EP3475858A1/fr
Publication of WO2018001847A1 publication Critical patent/WO2018001847A1/fr

Links

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H15/00—ICT specially adapted for medical reports, e.g. generation or transmission thereof
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
    • G06T17/10—Constructive solid geometry [CSG] using solid primitives, e.g. cylinders, cubes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0012—Biomedical image inspection
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H30/00—ICT specially adapted for the handling or processing of medical images
    • G16H30/20—ICT specially adapted for the handling or processing of medical images for handling medical images, e.g. DICOM, HL7 or PACS
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H30/00—ICT specially adapted for the handling or processing of medical images
    • G16H30/40—ICT specially adapted for the handling or processing of medical images for processing medical images, e.g. editing
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H40/00—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices
    • G16H40/60—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the operation of medical equipment or devices
    • G16H40/63—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the operation of medical equipment or devices for local operation
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30004—Biomedical image processing
    • G06T2207/30096—Tumor; Lesion

Definitions

  • CT computed tomography
  • MRI magnetic resonance imaging
  • PET positron emission tomography
  • the imaging commonly produces a stack of radiology images, which are sometimes referred to as image slices.
  • image slices In a common orientation, the patient lies in a prone or supine position, the image slices are transverse to the axial direction (running head-to-toe) of the patient, and the stack of radiology images is thus a stack of axial images.
  • a typical radiology reporting workflow involves the radiologist looking through the radiology images of the stack of radiology images acquired in an imaging study, identifying a radiology image of the stack which best depicts a finding for a tumor, lesion, or other feature of clinical interest, and dictating a narrative report describing the observed radiology findings.
  • a snapshot of a reference image chosen by the radiology may be saved on a Picture Archiving and Communication System (PACS) as the key image for future reference.
  • PPS Picture Archiving and Communication System
  • the radiologist may not make a careful selection of the reference image. This makes it difficult for the downstream consumers of the reports (e.g., referring physicians) to navigate to the images where the radiologist observed the findings and match them with the descriptions in the report.
  • the specific finding may correspond to visualization that spans through several slices and therefore, it may not be obvious to the radiologist which is the most appropriate image slice to be considered as the key image to match the finding. Further, there is inherent user variability, resulting in various inconsistencies.
  • a radiology workstation includes a computer connected to receive a stack of radiology images of a portion of a radiology examination subject.
  • the computer includes at least one display component and at least one user input component.
  • the computer includes at least one processor programmed to: display selected radiology images ofthe stack of radiology images on the at least one display component; receive entry of a current radiology report via the at least one user input component and displaying the entered radiology report on the at least one display component; identify a radiology finding by at least one of (i) automated analysis of the stack of radiology images and (ii) detecting textual description of the radiology finding in the radiology report; identify or extract at least one key image from the stack of radiology images depicting the radiology finding; and embed or link the at least one key image with the radiology report.
  • a radiology workstation includes a computer connected to receive a stack of radiology images of a portion of a radiology examination subject.
  • the computer includes at least one display component and at least one user input component.
  • the computer includes at least one processor programmed to: display selected radiology images ofthe stack of radiology images on the at least one display component; receive entry of a current radiology report via the at least one user input component and display the entered radiology report on the at least one display component; convert the stack of radiology images to a three-dimensional (3D) image; identify a radiology finding including a feature depicted in the 3D image; extract at least one two-dimensional (2D) key image depicting the radiology finding from the 3D image; and embed or link the at least one extracted 2D key image with the radiology report.
  • 3D three-dimensional
  • a radiology workstation includes a computer connected to receive a stack of radiology images of a portion of a radiology examination subject.
  • the computer includes at least one display component and at least one user input component.
  • the computer includes at least one processor programmed to: display selected radiology images ofthe stack of radiology images on the at least one display component; receive entry of a current radiology report via the at least one user input component and display the entered radiology report on the at least display component; detect identification via the at least one user input component of a location on a displayed selected radiology image of the stack of radiology images; identify a radiology finding for the feature at the identified location by detecting textual description of the radiology finding in the radiology report; identify or extract at least one key image from the stack of radiology images depicting the radiology finding; and embed or link the at least one key image with the radiology report.
  • One advantage resides linking an image with an identified feature to a corresponding radiology report.
  • Another advantage resides in automatically saving identified images having the identified feature.
  • Another advantage resides in providing concurrent and optionally visually linked display of text of a radiology report and an image having an identified feature so as to reduce likelihood of error in transferring content from the image to the radiology report.
  • FIGURE 1 shows a radiology workstation.
  • FIGURE 2 shows a flowchart showing an exemplary method of use of the device of FIGURE 1.
  • FIGURE 3 shows an example operation of the device of FIGURE 1.
  • the radiologist reviews a stack of 2D images (e.g., axial, sagittal, or coronal images in CT, MR, or SPECT).
  • the radiologist can flip through images of the stack to identify an axial image that best captures a feature of interest, such as a tumor or lesion.
  • Standard practice is that when a finding is made, for example a measurement of the size of the tumor or lesion, the radiologist may identify the key image in which the finding was observed/measured.
  • This key image is stored with the radiology report as a reduced-resolution snapshot image, possibly with annotations such as labeling the tumor and/or showing the line along which the tumor measurement was made.
  • the full-resolution DICOM image may also be hyperlinked with the key image, so that it can be easily pulled up when reviewing the report.
  • the radiologist has a short time interval to perform the reading, so that the key image selection may be cursory and the key image may not be ideal for making the finding.
  • a further difficulty is that radiologists conventionally review the images in their native format, such as the stack of axial slices.
  • the radiologist may be concerned that a non-native format may introduce image artifacts.
  • Time constraints also again come into play - a stack of N slices requires reviewing N images, whereas if the 3D image space is explored the number of potential images is infinite.
  • the tumor may have its largest and/or smallest dimension in some oblique or non-axial plane.
  • Improved radiology workstation embodiments disclosed herein provide automated selection of the "best" key image, and optionally also automated generation of an optimized finding in that best key image.
  • the key image selection module may be triggered automatically by detection via image processing of a tumor; or semi-manually by detecting a user interfacing operation such as using the cursor to define a tumor measurement line, or entry of the finding in the text of the report.
  • the activated module searches the stack of axial slices for the best axial slice for illustrating the finding. In another embodiment, the full 3D space is searched for the best slice of any orientation.
  • the user is presented with two or three key images: a native axial slice for which the tumor has largest area; an interpolated oblique slice showing the largest tumor dimension; and optionally also an interpolated oblique slice showing the smallest tumor dimension.
  • the module also automatically performs the finding computation for each identified key image.
  • the image processing identifies one or more tumors and key image(s) for each tumor, and the user is presented with the key image and the proposed finding for each tumor in turn.
  • the key image selection module upon detecting user entry of a finding, the key image selection module is invoked and identifies the best key image (and optionally the corresponding measurement) and proposes it to the radiologist.
  • the automated key image generation approaches disclosed herein can be extended to radiology findings other than tumor size findings, such as to lung texture findings indicative of emphysema, detecting calcifications in cardiac imaging, or so forth.
  • empirical feature selection and machine learning approaches can be used, i.e. a key image selector is trained on a set of labeled training imaging sets.
  • Illustrative embodiments automatically detect the most appropriate key image(s) corresponding to the finding description (based on the textual content in the report being created) to generate direct associations between report content and images.
  • the detected image(s) can also be automatically added as a key image to PACS.
  • a radiology reporting support application module for example implemented as a computer programmed to perform the disclosed operations, that automatically determines the radiology findings during report creation (for instance, during dictation) and automatically analyzes the image stack under review to determine the optimal key image that best describes the finding.
  • the device includes a radiology workstation 10, which may for example be implemented as a desktop computer, a "dumb" terminal connected with a network server, or any other suitable computing device.
  • the workstation 10 includes a computer 12 with typical components, such as at least one display component 14, a user input component 16, a communication link 18, a database 20 such as a Picture Archiving and Communication System (PACS), and at least one processor 22 programmed to perform radiology reporting functions as disclosed herein.
  • the at least one display 14 is configured to display one or more radiology images. For example, a selected image can be displayed on a first display, and the radiology report can be displayed on a second display.
  • the display 14 can be a touch-sensitive display.
  • the user input component 16 is configured to select at least one of the images.
  • the user input component 16 can be a mouse, a keyboard, a stylus, and the like.
  • the user input component 16 can be a microphone (i.e., to allow the user to dictate content to at least one of the radiology images), a headset, or simply be a user's finger (i.e., on a touch screen).
  • the communication link 18 can be a wireless communication link (such as a wired or wireless Ethernet link, and/or a WiFi link), e.g. a hospital network enabling the radiology workstation 10 to retrieve a stack of radiology images making up a study from the PACS 20.
  • the PACS or other database 20 is configured to store the stack of images, along with a plurality of radiology reports that include data entry fields indicative a type of the at least one target of interest.
  • the at least one processor 22 is programmed to cause the radiology workstation 10 to perform an automatic key image detection method, as described in more detail below.
  • the computer 12 is configured to receive a stack of radiology images of a portion (or all) of a radiology examination subject which are stored in the PACS 20. These images were previously uploaded to the PACS 20 from an imaging device (not shown) (e.g., a magnetic resonance device, an ultrasound device, a computer tomography device, a positron emission tomography device, a single-photon emission computed tomography device, and the like).
  • an imaging device not shown
  • a magnetic resonance device e.g., a magnetic resonance device, an ultrasound device, a computer tomography device, a positron emission tomography device, a single-photon emission computed tomography device, and the like.
  • the at least one processor 22 of the workstation 10 is programmed to cause the workstation 10 to perform an automatic key image detection method 100.
  • the method 100 includes: displaying selected radiology images of the stack of radiology images on the at least one display component 14 (102); receiving entry of a current radiology report via the at least one user input component 16 and displaying the entered radiology report on the at least display component (104); identify a radiology finding by at least one of (i) automated analysis of the stack of radiology images (106); and (ii) detecting textual description of the radiology finding in the current radiology report (108); identify or extract at least one key image from the stack of radiology images depicting the radiology finding (110); and embed or link the at least one key image with the current radiology report (112).
  • the identified feature is a tumor or a lesion.
  • the identified feature may be any suitable tissues (e.g., healthy tissue, a cyst, other abnormalities, tissue texture, and the like).
  • selected radiology images of the stack of radiology images are displayed on the at least one display component 14.
  • the at least one processor 22 is configured to retrieve a selected image (not shown), either by receiving the images directly from an imaging device (not shown), or from the database 20 (not shown), and display it on the at least one display 14.
  • a first image can be displayed on a first display 14, and a second image can be displayed on a second display (not shown).
  • the selected image can be displayed on the first display 14, and a radiology report retrieved from the database 20 can be displayed on the second display.
  • the image(s) is selected by a user (i.e., a medical professional such as a physician, a nurse, a technician, and the like) of the workstation 10.
  • a user i.e., a medical professional such as a physician, a nurse, a technician, and the like
  • entry of a current radiology report is received via the at least one user input component 16, and the entered radiology report is displayed on the at least display component 14.
  • the radiology reports are also stored in the database 20, along with the images.
  • the user may select at least one radiology report (e.g., by clicking on it or pointing to it with the user input component 16 (i.e., a mouse), by using text entry with the user input component 16 (i.e., a keyboard), or using dictation to select the radiology report with the user input component (i.e., a microphone), or may elect to create a new report, either ab initio or, more commonly, starting from a report template.
  • the report including content entered thus far
  • the report can be displayed on the display component 14 with the image (or on a separate display component).
  • a radiology finding is identified in the selected images.
  • the at least one processor 22 is programmed to analyze a subset of the image stack (i.e., an interval of images around the slice being viewed at the moment by the radiologist) or the whole image stack.
  • the radiology finding is identified by an automated analysis of the stack of radiology images. To do so, the at least one processor 22 is programmed to convert the displayed stack of radiology images to a three- dimensional (3D) image. The at least one processor 22 is then programmed to automatically identify a feature depicted in the 3D image.
  • the at least one processor 22 is programmed to execute one or more image processing algorithms (e.g., segmentation, atlas-based analysis, etc.) to determine the most appropriate slice that match the findings determined through text analysis, as described in more detail below.
  • image processing algorithms e.g., segmentation, atlas-based analysis, etc.
  • these automatic image processing algorithms for detection of the identified features can be prepared through machine learning algorithms and training-based algorithms using a database of training images with a known ground truth (e.g. radiology finding labels annotated to the images by skilled radiologists).
  • the radiology finding is identified by detecting a textual description of the radiology finding in the radiology report.
  • the at least one processor 22 is programmed to detect an identification via the at least one user input component 16 of a location on a displayed selected radiology image of the stack of radiology images. For example, the user can click on the displayed image (i.e., with the mouse) to identify a portion of the selected image (e.g., a dimension or axis of the region of interest), or use dictation (i.e., with the microphone) to identify a portion of the selected image (e.g., a dimension or axis of the region of interest).
  • the user can click on the displayed image with the mouse to define a length along the image.
  • the at least one processor 22 is then programmed to identify a feature at the identified location by automated analysis of the displayed selected radiology image of the stack of radiology images using the algorithms (e.g., segmentation, atlas-based analysis, etc.) described above.
  • the at least one processor 22 may convert the displayed stack of radiology images to a 3D image.
  • At 110 at least one key image is identified or extracted from the stack of radiology images depicting the radiology finding.
  • the key image is identified as having a desired or target criterion of the region of interest.
  • the radiology finding having desired or target criterion of the region of interest may constitute an identified feature in the selected image as having the largest cross-section.
  • the image in stack of images that the at least one processor 22 determines to have an identified feature with the largest cross-section is considered to be a key image.
  • more than one feature may be identified as having a larger cross-section relative to surrounding areas.
  • One difficulty with the approach of identifying the key image from the radiology images of the stack of radiology images is that the orientation of those "native" images may not be ideal for making or illustrating the finding. For example, if the imaging study acquired a stack of axial image slices and the finding is the long axis of a tumor, if the tumor is oriented such that its long axis does not lie in any axial plane then the finding cannot be effectively illustrated using the native (i.e. as-acquired) axial images.
  • the radiology finding having the desired or target criterion of the region of interest may constitute a maximum length of the identified feature along a long axis of the identified feature.
  • the stack of radiology images is converted to a three-dimensional (3D) image. This can be done, for example, by treating the two orthogonal dimensions of the image slices as the x- and y-coordinates of an x-y-z Cartesian space. If the spacing between slices is known (this spacing is commonly stored as metadata annotated to the images or to the image study) then the z-coordinate is defined orthogonal to the slices with data points (voxels) spaced apart along the z-direction by the inter-slice spacing.
  • the at least one key image includes a key image extracted from the 3D image that contains the long axis of the identified radiology feature (i.e., the maximum length of the identified feature along a long axis of the identified feature). Because of the three-dimensionality of the 3D image, it is always possible to find a (possibly oblique) plane that contains the long axis using standard Euclidean geometry principles. As a result, the image in stack of images (or the 3D image) that the at least one processor 22 determines to have an identified feature with the maximum length of the identified feature along a long axis of the identified feature is considered to be a key image.
  • the radiology finding having the desired or target criterion of the region of interest may constitute a minimum length of the identified feature along a short axis of the identified feature.
  • the at least one key image includes a key image extracted from the 3D image that contains the short axis of the identified feature.
  • the image in stack of images (or the 3D image) that the at least one processor 22 determines to have an identified feature a minimum length of the identified feature along a short axis of the identified feature is considered to be a key image.
  • more than one feature may be identified as having a minimum length of the identified feature along a short axis of the identified feature relative to surrounding areas.
  • any suitable desired or target criterion of the region of interest can be used to identify one or more key images.
  • an image corresponding to the mass centroid of the finding can be a key image.
  • a computational algorithm is used to determine the most specific body part and/or organ of the image where a measurement is made. Whenever a measurement is made on an image, this information is used to determine the body part and/or organ associated with this image (see Mabotuwana T, Qian Y, Determining scanned body part from DICOM study description for relevant prior study matching.., Stud Health Technol Inform., 2013; 192:67-71).
  • the radiology finding is identified by detecting textual description of the radiology finding in the radiology report.
  • the at least one processor 22 is programmed to analyze the radiology report.
  • the at least one processor 22 is programmed to monitor the text of the radiology report being entered can to detect entry of key words corresponding to standard radiology findings, e.g. using Systematized Nomenclature of Medicine (SNOMED) terminology optionally augmented by additional shorthand and/or alternative terminology in common use by radiologists.
  • SNOMED Systematized Nomenclature of Medicine
  • standard natural language processing (NLP) techniques can be used to extract findings in the current report.
  • the at least one processor 22 is programmed to analyze the radiology report with an NLP engine to determine sections, paragraphs, and sentences. The sentences are then extracted and processed by an image-based finding extraction engine executed by the at least one processor 22. For instance, this engine will extract sentence "The left hepatic reference lesion measures approximately 6.4 x 5.4 cm".
  • the at least one key image is embedded or linked with the radiology report.
  • the embedded or linking of the key image(s) to the radiology report is performed by a method 200 that is executed by the at least one processor 22.
  • the method 200 includes: retrieving at least one radiology report (202); detecting a finding in the at least one radiology report (204); automatically detecting at least one key image that includes the detected finding (206); saving the detected key image to a database (208); and embedded or linking the detected finding in the at least one key image to the radiology report (210).
  • At 202 at least one radiology report is created or retrieved.
  • the radiology reports can be created using a suitable report template, or a partially completed radiology report which is stored in the database 20 can be retrieved, along with the images.
  • the user may create or select at least one radiology report (e.g., by clicking on it or pointing to it with the user input component 16 (i.e., a mouse), by using text entry with the user input component 16 (i.e., a keyboard), or using dictation to select the radiology report with the user input component (i.e., a microphone).
  • the report can be displayed on the display component 14 with the image (or on a separate display component), and the radiologist can enter text reporting on radiology findings into the report, e.g. using a keyboard or a dictation microphone with speech recognition software running on the computer.
  • a finding is detected in the at least one radiology report.
  • the radiology finding is identified by detecting textual description of the radiology finding in the radiology report.
  • the at least one processor 22 is programmed to analyze the radiology report.
  • the at least one processor 22 is programmed to use standard natural language processing (NLP) techniques to extract findings in the current report.
  • NLP standard natural language processing
  • the at least one processor 22 is programmed to analyze the radiology report with an NLP engine to determine sections, paragraphs, and sentences. The sentences are then extracted and processed by an image-based finding extraction engine executed by the at least one processor 22.
  • NLP natural language processing
  • at least one key image that includes the detected finding is automatically detected. This process is described in more detail at 106, 108, and 110, described above.
  • the detected key images are saved to a database.
  • the key images are saved to the database 20.
  • the key images are saved to a PACS database (not shown). This can be performed using the communication link 18.
  • the detected finding in the key images is embedded or linked to the radiology report.
  • the at least one processor 22 is programmed to automatically convert the key image information (e.g., coordinates, slice number, and the like) into text and automatically add this text to the report (for example in the form of a hyperlink). To do so, the at least one processor 22 is programmed to translate image information (such as slice number coordinates) for uniquely identifying the key image. The at least one processor 22 is then programmed to insert the converted text into the appropriate spot in the report.
  • the embedded text can be in the form of a hyperlink. By clicking on hyperlink, the corresponding image study is opened and the image slice corresponding to the key image is shown on the display component 14.
  • the display component 14 is configured to populate the radiology report with the hyperlink containing the text content.
  • a hyperlink 30 containing the finding from a selected image 32 can be embedded into a radiology report 34 (shown by the dashed arrow in FIGURE 3).
  • the image 32 and the report 34 are displayed on the display component 14.
  • a communication network e.g., a wireless network, a local area network, a wide area network, a personal area network, BLUETOOTH®, and the like.
  • the various components 12, 14, 16, 18, 20, of the workstation 10 can include at least one microprocessor 22 programmed by firmware or software to perform the disclosed operations.
  • the microprocessor 22 is integral to the various components 12, 14, 16, 18, 20, so that the data processing is directly performed by the various components 12, 14, 16, 18, 20,.
  • the microprocessor 22 is separate from the various component v.
  • the data processing components 22 of the device 10 may also be implemented as a non-transitory storage medium storing instructions readable and executable by a microprocessor (e.g. as described above) to implement the disclosed operations.
  • the non-transitory storage medium may, for example, comprise a read-only memory (ROM), programmable read-only memory (PROM), flash memory, or other repository of firmware for the various components 12, 14, 16, 18, 20, and data processing components 22. Additionally or alternatively, the non-transitory storage medium may comprise a computer hard drive (suitable for computer- implemented embodiments), an optical disk (e.g. for installation on such a computer), a network server data storage (e.g. RAID array) from which the various component 12, 14, 16, 18, 20, data processing components 22, or a computer can download the device software or firmware via the Internet or another electronic data network, or so forth.
  • ROM read-only memory
  • PROM programmable read-only memory
  • flash memory or other repository of firmware for the various components 12, 14, 16, 18, 20, and data processing components 22.
  • the non-transitory storage medium may comprise a computer hard drive (suitable for computer- implemented embodiments), an optical disk (e.g. for installation on such a computer), a network server data storage (e.g. R

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • General Health & Medical Sciences (AREA)
  • Public Health (AREA)
  • Epidemiology (AREA)
  • Primary Health Care (AREA)
  • Radiology & Medical Imaging (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Physics & Mathematics (AREA)
  • Biomedical Technology (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Geometry (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Quality & Reliability (AREA)
  • Software Systems (AREA)
  • Computer Graphics (AREA)
  • Databases & Information Systems (AREA)
  • General Business, Economics & Management (AREA)
  • Business, Economics & Management (AREA)
  • Data Mining & Analysis (AREA)
  • Pathology (AREA)
  • Measuring And Recording Apparatus For Diagnosis (AREA)
  • Apparatus For Radiation Diagnosis (AREA)
  • Multimedia (AREA)

Abstract

Un poste de travail de radiologie (10) comprend un ordinateur (12) connecté pour recevoir une pile d'images radiologiques d'une partie d'un sujet d'examen radiologique. L'ordinateur comprend au moins un composant d'affichage (14) et au moins un composant d'entrée d'utilisateur (16). L'ordinateur comprend au moins un processeur (22) programmé pour : afficher les images radiologiques sélectionnées de la pile d'images radiologiques sur le ou les composants d'affichage ; recevoir l'entrée d'un rapport de radiologie actuel par le biais du ou des composants d'entrée d'utilisateur et afficher le rapport de radiologie entré sur le ou les composants d'affichage ; identifier un résultat radiologique à l'aide d'au moins l'un des éléments suivants : (i) l'analyse automatique de la pile d'images radiologiques et (ii) la détection de la description textuelle du résultat radiologique dans le rapport de radiologie ; identifier ou extraire au moins une image clé de la pile d'images radiologiques décrivant le résultat radiologique ; et intégrer ou lier la ou les images clés au rapport de radiologie.
PCT/EP2017/065347 2016-06-28 2017-06-22 Système et procédé de détection automatiques d'images clés WO2018001847A1 (fr)

Priority Applications (3)

Application Number Priority Date Filing Date Title
CN201780046678.5A CN109564773B (zh) 2016-06-28 2017-06-22 用于自动检测关键图像的系统和方法
US16/314,104 US11361530B2 (en) 2016-06-28 2017-06-22 System and method for automatic detection of key images
EP17735024.6A EP3475858A1 (fr) 2016-06-28 2017-06-22 Système et procédé de détection automatiques d'images clés

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US201662355411P 2016-06-28 2016-06-28
US62/355,411 2016-06-28

Publications (1)

Publication Number Publication Date
WO2018001847A1 true WO2018001847A1 (fr) 2018-01-04

Family

ID=59276699

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/EP2017/065347 WO2018001847A1 (fr) 2016-06-28 2017-06-22 Système et procédé de détection automatiques d'images clés

Country Status (4)

Country Link
US (1) US11361530B2 (fr)
EP (1) EP3475858A1 (fr)
CN (1) CN109564773B (fr)
WO (1) WO2018001847A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2022152733A1 (fr) * 2021-01-15 2022-07-21 Koninklijke Philips N.V. Post-traitement d'images radiologiques

Families Citing this family (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016179310A1 (fr) * 2015-05-04 2016-11-10 Smith Andrew Dennis Estimation de réponse tumorale assistée par ordinateur et évaluation de la charge tumorale vasculaire
DE112020005870T5 (de) * 2019-11-29 2022-11-03 Fujifilm Corporation Unterstützungsvorrichtung für dokumentenerstellung, unterstützungsverfahren für dokumentenerstellung und unterstützungsprogramm für dokumentenerstellung
JPWO2021157705A1 (fr) * 2020-02-07 2021-08-12
US20210398285A1 (en) * 2020-06-17 2021-12-23 Fovia, Inc. Consecutive slice finding grouping
US20210398277A1 (en) * 2020-06-17 2021-12-23 Fovia, Inc. No results indicator for stack of images
US20210398653A1 (en) * 2020-06-17 2021-12-23 Fovia, Inc. Key image updating multiple stacks
US11580337B2 (en) * 2020-11-16 2023-02-14 International Business Machines Corporation Medical object detection and identification
US12106850B2 (en) * 2021-01-28 2024-10-01 RAD AI, Inc. System and method for workflow management and image review
WO2023205181A1 (fr) * 2022-04-19 2023-10-26 Synthesis Health Inc. Création d'images composites à l'aide d'une compréhension de langage naturel

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090309874A1 (en) * 2008-06-11 2009-12-17 Siemens Medical Solutions Usa, Inc. Method for Display of Pre-Rendered Computer Aided Diagnosis Results
US20140257854A1 (en) * 2011-09-08 2014-09-11 Radlogics, Inc. Methods and Systems for Analyzing and Reporting Medical Images
WO2015092633A1 (fr) * 2013-12-20 2015-06-25 Koninklijke Philips N.V. Création automatique d'une vue longitudinale centrique de constations concernant un patient

Family Cites Families (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7929743B2 (en) * 2007-10-02 2011-04-19 Hologic, Inc. Displaying breast tomosynthesis computer-aided detection results
JP5226370B2 (ja) * 2008-04-18 2013-07-03 株式会社東芝 読影レポート作成支援システム
US8170310B2 (en) * 2009-02-23 2012-05-01 General Electric Company Automatic outcome analysis using radiological images
JP5455470B2 (ja) 2009-07-02 2014-03-26 株式会社東芝 医用画像読影システム
WO2012071016A1 (fr) * 2010-11-26 2012-05-31 Agency For Science, Technology And Research Procédé de création de rapports radiologiques multimédia
US20140006926A1 (en) 2012-06-29 2014-01-02 Vijaykalyan Yeluri Systems and methods for natural language processing to provide smart links in radiology reports
US8935155B2 (en) 2012-09-14 2015-01-13 Siemens Aktiengesellschaft Method for processing medical reports
US20140149942A1 (en) * 2012-11-23 2014-05-29 Cleon Hill Wood-Salomon System and method for actionable reporting of a radiology image study
US9904966B2 (en) * 2013-03-14 2018-02-27 Koninklijke Philips N.V. Using image references in radiology reports to support report-to-image navigation
US10657220B2 (en) * 2015-04-22 2020-05-19 Ascend Hit Llc System and methods for medical reporting

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090309874A1 (en) * 2008-06-11 2009-12-17 Siemens Medical Solutions Usa, Inc. Method for Display of Pre-Rendered Computer Aided Diagnosis Results
US20140257854A1 (en) * 2011-09-08 2014-09-11 Radlogics, Inc. Methods and Systems for Analyzing and Reporting Medical Images
WO2015092633A1 (fr) * 2013-12-20 2015-06-25 Koninklijke Philips N.V. Création automatique d'une vue longitudinale centrique de constations concernant un patient

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
MABOTUWANA T; QIAN Y: "Determining scanned body part from DICOM study description for relevant prior study matching.", STUD HEALTH TECHNOL INFORM., vol. 192, 2013, pages 67 - 71
SHIN HOO-CHANG ET AL: "Interleaved text/image Deep Mining on a large-scale radiology database", 2015 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), IEEE, 7 June 2015 (2015-06-07), pages 1090 - 1099, XP032793561, DOI: 10.1109/CVPR.2015.7298712 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2022152733A1 (fr) * 2021-01-15 2022-07-21 Koninklijke Philips N.V. Post-traitement d'images radiologiques

Also Published As

Publication number Publication date
EP3475858A1 (fr) 2019-05-01
CN109564773A (zh) 2019-04-02
US20190325249A1 (en) 2019-10-24
CN109564773B (zh) 2023-07-07
US11361530B2 (en) 2022-06-14

Similar Documents

Publication Publication Date Title
US11361530B2 (en) System and method for automatic detection of key images
JP6749835B2 (ja) コンテキスト依存医学データ入力システム
EP2888686B1 (fr) Détection et récupération automatiques d'annotations antérieures pertinentes pour une étude d'imagerie en vue d'une visualisation et d'un établissement de rapport d'une grande efficacité
AU2006254689B2 (en) System and method of computer-aided detection
US10497157B2 (en) Grouping image annotations
EP2881916B1 (fr) Génération spécifique de demande et récupération d'images de volume médicale
US10210310B2 (en) Picture archiving system with text-image linking based on text recognition
JP7258772B2 (ja) 全人的患者放射線医療ビューワ
US11468567B2 (en) Display of medical image data
EP3191991B1 (fr) Identification d'annotation de rapport d'image
US20170221204A1 (en) Overlay Of Findings On Image Data
US10614335B2 (en) Matching of findings between imaging data sets
US10235360B2 (en) Generation of pictorial reporting diagrams of lesions in anatomical structures
JP2024528381A (ja) 医用画像データの自動追跡読影方法及びシステム
EP4310854A1 (fr) Dispositif, procédé et programme de création de document
CN117581310A (zh) 用于医学图像数据的自动跟踪读取的方法和系统
US20120191720A1 (en) Retrieving radiological studies using an image-based query
JPWO2023274599A5 (fr)
WO2014072928A1 (fr) Facilitation d'interprétation d'une image médicale

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17735024

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2017735024

Country of ref document: EP

Effective date: 20190128